The Nikon D850 seems fine. Almost all profiles are good. This specific Canon was the worst case. The Leica M9 has a slight color shift. And there was a Pentax model with a shift, I need to check which one later. Also most of the film simulation profiles have color shifts (not all), but I guess that is intended.
A long story, but sorry, my question is not answered yet.
Why are all WB data relative between brand, camera and software, but not between these all and âdaylightâ PhotoLab namely 5200 K and Tint 0 ?
I think this is a design decision DxO made. All white balance settings, except the as shot value, are fixed values, âdaylightâ is 5200 K, âcloudyâ is 6000 K, etc., always with tint 0.
I think that this can be problematic. Imagine, you have a white balanced image, where Photolab will consider the tint as -15. All other white balance modes will basically be useless, as that tint is not considered.
Capture One might have calibrated their white balance settings for each camera model individually. This is obviously not the case for Photolab.
On the other hand, from a historic (film) perspective, âdaylightâ has always been 5600°K and that is what I always set my camera to rather than what anyone deems to be daylight nowadays. I believe the change came about because a lot of folks donât process their images and 5200°K gives a warmer, âfriendlierâ rendering in that instance.
Although, from my findings, not all cameras stick to 5200°K and âdaylightâ can vary.
But despite setting the temperature manually, PL still seems to do what it thinks rather than simply honouring what I want.
This is a clear answer, thank you!
Most colour films were balanced to 12 oâclock noon day sunshine on a clear day, which was considered to be 5600 K. Like Joanna, both my cameras are set to 5600 K and my default in Photolab is also set to 5600 K. Most artificial light films were set to 3200 K and if I remember correctly, a north facing window at 12 oâclock on a clear sunny day was considered to be 6000 K. I do stand to be corrected as memory is getting a bit thin on the ground.
Prem.
From nature prespective daylight is depening on what coordinates on earth you are. And which time of year. And ofcoarse time of day.
We have 5200-5600K adopted as Daylight color.
In reality we have lots of different daylight colors and thus expected WB correction to 5600K.( our eyeâs and brains does this automatic.)
To be measuring as correct as possible i think you need date, time(astronomical timetable), coordinates on earth where you take the image and color sensitive correctiontable of the used sensor and circuit. (aka whatâs white for the camera)
Too much trouble i think.
In this perspective all brands just gues what would be suitable for most âdaylightâ images and blend in images in one go for more then one camerabrand is even more troublesome. 
Nevertheless if they test bodyâs and lenscombinations in a lab with controlled light in order to get a baseline of WB and daylight definition it should be reasonable the same outcome between tested bodyâs and lenscombination.
Personally the DxO WB toolkit isnât very comprehencive. Things Like uniWB and a good Automated WB colorcast correction based on the image. Or what i would like a more detailed presetlist of WB or a personal correction of that list based on âdaylightdefinitionâ
If i say daylight is 5500K and default is 5400K every thing shifts 100K on the preset list.
This way every one can personalise the WB computing.
2 Likes